Can Diseases Be Managed to Cost Less?

Chronic diseases account for the vast majority of healthcare expenditures in the United States. Since these diseases are often preventable and frequently manageable through early detection, improved diet, exercise, and treatment therapy, a number of insurance companies have launched disease management programs to help members better manage their chronic diseases in an effort to reduce healthcare costs. Ideally, if patients have learnedto take better care of their health problems, there would be less healthcare service use and costs associated with avoidable complications. However, there is mixed evidence thatdisease managementprogramssave money. Indeed, in terms of highmanagement costsand various risk profiles of patients, it is not obvious that disease managementprograms reduce healthcare costs.Using private membership and claims data, this paper examines the effect of a diabetic disease program launched by an insurer company in Louisiana on healthcare costs. Through comparing the changes in healthcare spending before and after participation for members who participate in the program to those for members who do not participate, this paper identifies the plausibly causal effect. Considering the potential selection bias caused by voluntary participation of the program, I calculate propensity scores based on members’ health risk scores and other characteristics to match nonparticipants with participants. Given that participants are assigned different levels of intervention by health risk scores, I also use the regression discontinuity design to exploit plausibly random variation around the different stratification criteria, providing evidence on the heterogeneous effect by the intensity of the interventions. Additionally, this paper further exploits how this disease program affects healthcare costs through changing Emergency Room visits, hospitalization days and medication adherence.